New answers tagged mfcc
0
It just lift the high frequency part coefficients
we can refer the HTK book, equ 5.12
We can plot the 12 coeffs while using these condition:
N=13; %N is the number of cepstral coefficients
L=22; % L is the liftering parameter
ceplifter = ( 1+0.5*L*sin(pi*[0:N-1]/L) );
plot(ceplifter);
xlabel('number of cepstral coefficients');
ylabel('ceplifter');
...
Top 50 recent answers are included
Related Tags
mfcc × 174speech-recognition × 47
audio × 35
speech-processing × 25
fft × 19
machine-learning × 15
matlab × 14
signal-analysis × 13
python × 11
cepstral-analysis × 11
feature-extraction × 10
speech × 9
filters × 7
discrete-signals × 7
sound × 7
classification × 7
voice × 7
spectrogram × 6
dct × 6
sound-recognition × 6
fourier-transform × 5
hmm × 5
frequency-spectrum × 4
short-time-ft × 4
audio-processing × 4